# Reading your results # Reading your results A result is a historical window whose shape is similar to your pattern. This page explains what each number means and how to use the chart view. ## The results list Each row is one coin, keeping its best match: | Field | Meaning | |---|---| | **Symbol** | The coin and pair, for example `BTCUSDT` | | **Distance** | How far the segment is from your pattern. Lower is closer; it is normalised, so distances are comparable between coins | | **Score** | The combined quality score, higher is better. It blends the shape distance with the other checks you enabled | | **Match count** | How many separate windows in this coin matched | | **Direction** | Whether the matched window moved up or down overall | Sorting is by score. Use the filter box in the Results tab to narrow the list to a symbol. ## The chart Selecting a coin draws its price series with every matching segment highlighted, plus the first match overlay so you can compare it directly with your pattern. Zoom in to judge a segment up close before you trust it. ## The report A search also produces a report: one page with every coin's best match and its numbers, reachable by its own link. Reports are what you share with a teammate, and they are stored with your search history so you can come back to them. ## Statistics, honestly Numbers such as the historical win-rate figure are computed from past data over a limited sample. They describe what happened, they do not forecast what will happen, and they ignore fees, spread and slippage. Read the [Disclaimer](/docs/disclaimer) before acting on a match. ## Turning knobs that change what you get - **Tolerance (max distance)** — raise it to see more, weaker matches; lower it for fewer, closer ones. - **Period** — a longer scan sees more history and costs more. - **Filters** — turning on the angle, ratio, turn or peak checks rejects segments that only look similar at a glance. - **Bar range** — limits how long a candidate window may be, which speeds up long scans. ## No matches? That is a valid result. Tighten the pattern (fewer, clearer moves), raise the tolerance, widen the period, or switch to a warping algorithm such as **DTW** that tolerates different timing.